欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
现有的图像分块方法在处理前将源噪声图像分割成不同的小块，分别独立进行去噪，最后将所有小块拼接。该方法所得图像的去噪效果有一定程度的提高，但在相邻分块结合处存在不连续。对去噪方法的现状进行了分析，并结合最小化总变差及分块思想，提出新的去噪方法：在每小块处理过程中，分块值参与判断终止，整幅图像值计算去噪图像；对各块所得去噪图像取均值，得到最终的结果；对新算法进行优化，提高了算法运行效率。分析实验结果表明，该方法在计算效率和信噪比上均有显著提高。
英文摘要：
The existing image block method is tried to divide the whole image into blocks, then deals with them separately. The result is the combination of these denoised blocks. The denoising effect could be improved to a certain degree, but some discontinuities can be found in the areas of adjacent blocks. The current situation and characteristic of image denoising are reviewed, and a new algorithm is introduced, which is based on the minimizing total variation together with the image dividing. When calculating, the values of each block are used to judge whether it＇ s the time to stop or not, the values of the whole image are used to calculate the denoised image. Then the medium value of all denoised images which created by the blocks is the final result. The operational efficiency is improved obviously after the optimizing of the algorithm. With the analysis of the experimental data, this method reaches a good result both on the execute time and the SNR.
同期刊论文项目
同项目期刊论文
Stability analysis for the generalized Hopfield neural networks with multi-level activation function
期刊信息
位置：